This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. As described in earlier sections, the collaborators on this initiative have proposed challenging structural problems that demand solution SAXS and concomitant advances in phasing methods. A good example is Joseph Wedekind's (U. Rochester) investigation of the structural basis for function of the APOBEC[unreadable]3G (A3G) enzyme in human white blood cells, where it acts to stave off HIV infections. The team is making progress on crystallizing A3G, but needs complementary structural information with the ultimate goal of learning how this important protein responds to regulatory factors that target it for degradation. Wedekind has teamed up with Richard Gillilan to obtain the general shape of a molecule in solution utilizing SAXS. If and when crystals become available, the solution shape information will be used to phase the crystallographic structures. The proposed phasing procedure requires three steps: (1) Obtain a solution SAXS pattern and determine the molecular envelope. (2) If crystals exist, collect a standard crystallographic data set and use the envelope information for low-resolution phases. (3) Extend phases to higher resolution to solve for the high resolution structure. The procedures are detailed below. Obtain a SAXS Pattern and Determine the Molecular Envelope We propose to optimize experimental techniques to determine molecular envelopes using SAXS. The current setup for SAXS experiments at the CHESS G1 beamline is as follows (Gillilan, 2002, unpublished; Fig. 31): + G1 wiggler line with multilayer optics + 1.2 Angstrom wavelength with ~2% bandpass + Quantum4 2K (ADSC) detector @ ~780mm from sample cell + 500[unreadable]m x 500[unreadable]m beam with 0.8mm guard slits + He flight tube with Be window on sample end, 0.5mil mylar on detector end + 1mm path (80[unreadable]l) sample cell (courtesy of J.G. Grossmann, Daresbury U.K.) + 25[unreadable]m mica windows in sample cell Scattering Data Acquisition and Analysis Samples are encapsulated inside a cell sandwiched by two thin parallel windows. SAXS data of buffer and samples at different concentrations are collected. SAXS data are processed using the software developed by Xinguo Hong and Richard Gillilan [108]. The data reduction includes normalization of the scattered data to the intensity of the transmitted beam and subtraction of the background scattering of the buffer. All scattering curves are then standardized to that of a protein concentration of 1 mg/ml. The low angle data will be extrapolated to infinite dilution and merged with the high angle data measured at high protein concentrations to yield final scattering curves. Once the solution scattering data from a protein sample is obtained, the next step is to recover the three-dimensional envelope from the one-dimensional scattering pattern. Two methods developed by Svergun and colleagues will be used. In the first general ab initio approach [74,109], an angular envelope function of the particle, R = F( ), where (r, ) are spherical coordinates, is described by a series of spherical harmonics. The lowresolution shape is thus defined by a few parameters [unreadable]the coefficients of this series [unreadable]that fit the scattering data. This approach was implemented in the computer program SASHA [75]. It was demonstrated that, under certain circumstances, a unique envelope can be extracted from the scattering data (except for the handedness of the shape [unreadable]this ambiguity holds for all ab initio methods in SAXS). Both `left` and `right` hands should be tested and the ambiguity resolved in the later stages of the structure determination, when the hand of helices can be determined. The use of envelopes defined by spherical harmonics is limited to globular particles with relatively simple shapes and without significant internal cavities. More detailed models can be constructed ab initio using different types of Monte Carlo searches and the utilization of a simulated annealing approach in which the shape of the complex is modeled by a large number of close-packed beads, which are moved around so as to match the observed scattering profile as well as possible [110-112]. The Monte-Carlo-based models contain hundreds or thousands of parameters, and caution is required to avoid over-interpretation. A common approach is to align a set of models resulting from independent shape reconstruction runs to obtain an average model that retains the most persistent, and presumably also most reliable, features, e.g. using the program SUPCOMB [113]. Particle symmetry, if known, provides very useful constraints, which can be imposed in the programs SASHA and DAMMIN, and in the program GASBOR [110-112]. Once the molecular shape is determined from the SAXS pattern and crystallographic data are collected, the molecular replacement method implemented in the FSEARCH program [93] is used to locate the envelope in the crystallographic unit cell, thus providing low resolution phases. FSEARCH (distributed by CCP4) can accept either of the two forms of envelope (spherical harmonics or Monte-Carlo-based models) as an input search model. It has been shown that the absence of strong reflections at low resolution caused by saturation at the detector can degrade FSEARCH solutions greatly [114]. Therefore, particular attention is paid in crystallographic data collection experiments to ensure low resolution data (100-10[unreadable]) are near complete (by using a small beam stop) and not saturated (by reducing exposure time).